The expanding universe of ribonucleoproteins: of novel RNA-binding proteins and unconventional interactions

Abstract

Post-transcriptional regulation of gene expression plays a critical role in almost all cellular processes. Regulation occurs mostly by RNA-binding proteins (RBPs) that recognise RNA elements and form ribonucleoproteins (RNPs) to control RNA metabolism from synthesis to decay. Recently, the repertoire of RBPs was significantly expanded owing to methodological advances such as RNA interactome capture. The newly identified RNA binders are involved in diverse biological processes and belong to a broad spectrum of protein families, many of them exhibiting enzymatic activities. This suggests the existence of an extensive crosstalk between RNA biology and other, in principle unrelated, cell functions such as intermediary metabolism. Unexpectedly, hundreds of new RBPs do not contain identifiable RNA-binding domains (RBDs), raising the question of how they interact with RNA. Despite the many functions that have been attributed to RNA, our understanding of RNPs is still mostly governed by a rather protein-centric view, leading to the idea that proteins have evolved to bind to and regulate RNA and not vice versa. However, RNPs formed by an RNA-driven interaction mechanism (RNA-determined RNPs) are abundant and offer an alternative explanation for the surprising lack of classical RBDs in many RNA-interacting proteins. Moreover, RNAs can act as scaffolds to orchestrate and organise protein networks and directly control their activity, suggesting that nucleic acids might play an important regulatory role in many cellular processes, including metabolism.

Keywords: Interactome capture; Protein-RNA interaction; RNA binding domain; RNA-binding proteins; RNA-determined RNPs; Ribonucleoproteins.

Fig. 1

RNA interactome capture discovers many RNA-binding proteins that lack identifiable RNA-binding domains. RNA interactome capture from different human cell lines [8, 13, 18] identified a total of 1218 proteins as RNA binders, most of which do not contain an identifiable RBD (∼55 %). The remaining proteins harbour domains known to bind RNA, most commonly the RNA recognition motif (RRM, accounting for ∼13 % of the proteins), DEAD/DEAD box helicase domain (accounting for ∼4 %) and the K homology domain (KH, accounting for ∼3 %)

Fig. 2

Protein-RNA interaction through low-complexity, extended protein regions and protein-binding RNAs. a, b The S. cerevisiae ribosomal protein L23 (RPL23) contains a C-terminal domain that folds into an RRM-like topology, whereas the N-terminal, low-complexity region adopts an idiosyncratic, extended conformation (a). In the 60S ribosomal subunit, the lysine- and arginine-rich N-terminal domain of RPL23 participates in extensive interactions with the ribosomal 25S and 5.8S RNAs (b). RNAs are depicted in grey, neighbouring proteins RPL8 and RPL35 in light blue; in the structure based on Ben-Shem et al. [14] (PDB ID: 4V88), residues within 20A of the RPL23 protein are depicted. c HCV IRES bound to a 40S ribosomal subunit. The HCV IRES (blue) displays an elongated structure that binds the solvent side of the 40S ribosomal subunit. Interactions are formed mostly with ribosomal proteins (eS1/S3A, uS7/S5, eS7/S7, uS11/S14, eS25/S25, eS26/S26, eS27/S27 and eS28/S28, proteins depicted in yellow) and, only to a lesser extent, with the ribosomal 18S RNA (shown in grey). In the structure based on Yamamoto et al. [123] (PDB ID: 5FLX), individual panels represent different orientations (rotated by 90°)